First software family to run on the grid

Zatni Arbi , Columnist, Jakarta
zatni@cbn.net.id

Suppose you are throwing a party at your house to celebrate your golden anniversary. You are going to invite all your relatives, friends, colleagues, neighbors -- what have you.

You will rent a powerful sound system, put in dozens of high- power spotlights, install a number of portable air conditioners in the house, and order a number of electric stoves to prepare food for hundreds of guests. Just imagine how much electricity you'll need for just that one night?

How are you going to meet the need? Are you going to invest in a multikilowatt generator just for one night of fun? If you are, what are you going to do with it until you need it again -- perhaps when your daughter's wedding day arrives 10 years down the road?

At least, as far as electricity goes, especially if you live in countries where electricity is largely taken for granted, you will have nothing to worry about. You just use the electricity as you need, and the electrical grid will meet your demand.

The word "grid", at least in the U.S. recently, has been a very familiar word. It became the word on everybody's lips on Aug. 14, when suddenly New Yorkers and people in the neighboring states found themselves without electricity.

The grid, in this case, referred to the interconnected systems that supply electricity to subscribers. There are many different sources of electricity made available on the grid. Some of the electricity may be generated by turbine-based generators, some may come from nuclear plants. And some may also be supplied by generators driven by diesel motors. Some power plants use steam to propel the generators, and others use coal.

A key aspect of the grid, needless to say, is the standards. Here in Indonesia, we may not have stringent standards -- which is why people in some places have stable currents while others have to use voltage stabilizers to maintain the voltage as close to the required 220 volts as possible. In the U.S., as well as other developed countries, no stabilizer is generally necessary, as the entire grid supplies the standard 110 or 220 volts.

Another key aspect of the grid is that it is supposed to be able to supply all the electricity that we need at any given time. So if you live in an apartment in San Francisco, you should be able to use as much electricity as you like as long as you pay for it. That has not always been the case, however, as the California power crisis demonstrated.

Yet, those who live in San Francisco are better off than, for example, the people who live in my neighborhood. Not only is our voltage generally far below the standard 220 volts, we cannot freely increase our electricity use for portable air conditioners when we are having a large birthday party, for example, as the power company PLN limits how much electricity we can use in our individual households. Of course, the reason for imposing the limits is to enable PLN to charge us a higher rate the more electricity we need daily.

A power grid, nonetheless, makes it possible for power companies to share resources. Anyone who needs electricity for his electric shaver or welding machine will only need to plug his device into the power outlet. He would not have to know where the electricity comes from, who makes it available, etc.

Grid Computing

Using the power grid as a model, researchers came up with the idea of grid computing. Instead of using one single, stand-alone computer with immense computing power and an extremely high price tag in isolation to run computer applications, why not connect a large number of smaller and cheaper computers together to do the same job?

It would be cheaper and it would use the computational cycles of the computers that would otherwise remain idle during the time they had anything to do. That is one of the basic ideas behind the computational grid. The idea itself emerged in the early 1970s, when computers began to be connected together in networks in several research centers in the U.S.

These researchers actually began building the computing grid in the mid-1990s when all the necessary components for the infrastructure became largely available -- Ethernet, Internet, Storage Area Network (SAN), Network Attached Storage (NAS), etc. -- to tackle heavy-duty computational jobs.

One of the examples of this computational power grid is Distributed.net. These grids have acquired various names, including utility computing (obviously because they enable us to use computing power much like we use electricity), on-demand computing, 10g, the grid for business

For businesses, the benefits of a computing grid are numerous, and they were highlighted by Larry Ellison, the chairman and CEO of Oracle Corp., during his speech at the recent OracleWorld 2003 in San Francisco. During this event, the company officially introduced its Oracle 10g, which is meant to run on an enterprise grid.

"We use the term 'enterprise grid' to differentiate what we mean from the grid used in the academic world," said Ellison at the beginning of his speech.

In my next article, we will take a closer look at some of the new features of the Oracle 10g family, which includes Oracle Database 10g and Oracle Application Server 10g. We will also explore the benefits of the grid to businesses, as promised by this largely welcomed update of Oracle products.

One thing to note is that the Oracle 10g, which everyone at the Redwood Shores-based company is so upbeat about, is the first enterprise software built to run on a grid. It does not mean that Oracle is the only company driving the use of computing grid into the business world, however. IBM, HP and Sun Microsystems have also been pushing the same idea.

However, for now, Oracle has beaten its competitors in the database arena -- IBM and Microsoft -- by adding the grid capability to its database product, making it more suitable to companies with large distributed data centers and computers that are idle much of the time, plus an urge to maximize the efficiency of their investments.